This study developed an early warning model for Korea’s salmon import supply chain, which relies heavily on a single country. A supply chain crisis is defined as a significant change in the CIF import price beyond a stable range, with potential impacts on domestic prices. The crisis index, using January 2010 as the base point, combines the relative price level and its year-on-year growth rate. The threshold was set based on earlier agricultural early warning studies. Monthly and quarterly data were used to select explanatory variables including supply factors and demand or macroeconomic factors. Variables with high 3~5 month lagged correlations were chosen using a stepwise method. Ordinary Least Squares (OLS) regression and Probit models were applied for both all crises and continuous crises, and predictive accuracy was evaluated using MAE and RMSE. Results show that the Probit model with a five-month lag for continuous crises provided the highest accuracy.